PrismML released the Ternary Bonsai 2 series with 2B, 4B, and 9B models built on Alibaba's Qwen 3.5, using ternary weights (-1, 0, +1) for on-device AI agents across phones, PCs, and wearables. Storage ranges from about 0.43GB to 2GB; the 9B model cuts token usage by ~40% in instruction-following and math tasks; the 2B model hits up to 120 tokens/s on iPhone 18 Pro Max. Weights are open under Apache 2.0.
PrismML released its latest reasoning model, Bonsai 2 27B, on Thursday. The model compresses Alibaba's open-source Qwen3.8 27B to 5.9GB, making it runnable on PCs and high-end smartphones, with a 9-10x reduction in memory usage and benchmark scores reaching 98% of the original model.